moon just fitted into the aperture of the V, while opposite sides of the
corona were reflected through the prisms to the place where they came
together. In this way both sides of the corona were seen through the
eye-piece at the same time. On looking at the eclipse this is what Dr.
Hastings saw: The light of the corona was divided into its constituents.
Prominent among them was a bright green line, which is designated by the
number 1,474; to this line attention was directed. Its presence in the
spectrum has been an argument in favor of the view that the corona is
a solar atmosphere. If this is the case, the line should remain fixed
during the eclipse; but if the corona is due to diffraction, this line
should change. It should grow shorter in the light from one side of the
corona, and longer on the other. The observation was now reduced to
watching for a change in the relative length of two green lines.

At the beginning of totality the line from the west side was much the
longer, but as the eclipse progressed it shortened notably, while the
line from the east side, shorter by about one-third at the beginning of
the eclipse, grew longer. When the eclipse ended, the proportions of the
lines were exactly reversed. There had been a change equal to two-thirds
the length of the lines, while the sun and moon had only changed their
relative positions by an extremely small amount. The only way in which
this phenomenon can be accounted for is on the diffraction theory. The
material view of the corona will not answer for it. But there are other
discrepancies in the older view which have been known for some time.
The principal ones are: 1. It is known from study of the sun that the
gaseous pressure at the surface must be less than an inch of mercury,
and is probably less than one-tenth of an inch, but an atmosphere
extending to the supposed limits would cause an enormous pressure at the
sun's surface, especially since the force of gravity on the sun is very
much greater than on the earth. 2. The laws of gravitation would require